NA Digest Sunday, January 29, 1995 Volume 95 : Issue 05

Today's Editor:
Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Michael Overton <>
Date: Sun, 22 Jan 95 16:15:07 EST
Subject: Notes on Numerical Computing

Over the last few years I have developed a set of notes called
"Notes on Numerical Computing", for use in place of a text in an
undergraduate numerical computing class intended for computer science
students (rather than math or engineering students).

The philosophy of these notes is to teach a small number of key ideas
in detail, instead of a lot of material. Topics which are emphasised are
the IEEE floating point standard and systems of linear equations, with
additional material on nonlinear equations, integration and ODE's.
The floating point standard material in particular contains much information
that is hard to find elsewhere. Everything is discussed at a low level,
without assuming much prior knowledge. The course is based on PC's,
using mostly Matlab but also some C and a special simulator which
demos many features of the IEEE standard. There is also some
optional material using the MacMath package (which runs only on Macs).

These notes have been used extensively at NYU, and also at McGill, where
Chris Paige made many excellent improvements which have been incorporated.
Years ago many CS depts had a required numerical computing class.
NYU and McGill still do, but most other depts have long since dropped the
requirement. Why? Somehow CS depts (students and faculty) got the idea
than numerical analysis is too hard, or boring, or at any rate not
mainstream computer science. These notes attempt to go the other way:
the material is easily accessible to the average CS undergraduate; and is,
we argue at least, both interesting and important. Student feedback has
been quite favorable.

The notes are reasonably polished but not at a publishable stage. If you
would like to try them in a course you are teaching, and are willing to
give detailed feedback from your experience with them, please send info on
the course title, dates and enrollment.

Michael Overton.


From: Andrew Knyazev <>
Date: Sun, 22 Jan 1995 20:18:03 -0700
Subject: Promised "New Estimates for Ritz Vectors" by FTP

Dear Netters,

At a linear algebra conference last summer I promised several persons
to send a text of my talk. Unfortunately, I have lost their addresses then.
The text, "New estimates for Ritz vectors," has been recently published
as a Tech. Report #677 of Dept. Comp. Sci. at NYU's CIMS, and is available
at our anonymous ftp cite, or using URL:
as (41K).

The purpose is to make clear a behavior of Ritz vectors ...,e.g.\
for a case when the trial subspace is close to an invariant subspace...
Theorem 3.1 shows that the approximation error of an eigenvector
by the corresponding Ritz vector is essentially orthogonal to this
invariant subspace. Theorem 3.2 leads to a dual statement that
the orthoprojection of any eigenvector from the invariant subspace
onto the trial subspace essentially coincides with a Ritz vector.

There is also my eigenproblem-oriented bibliography file
eigenproblem.bib.gz (15K) in the same directory.

Thanks for your attention, Andrew Knyazev


From: Fred Kus <>
Date: Tue, 24 Jan 1995 10:29:34 -0500
Subject: Computing the Jordan Normal Form


I am looking for a program to calculate the Jordan normal form of a real
or complex matrix. The only routine I have seen is TOMS 560. Is there anything
more recent, which uses LAPACK and BLAS routines ?

Thanks in advance.


Fred W. Kus INTERNET: fred@mcmail.CIS.McMaster.CA
Computing & Information PHONE: (905) 525-9140 ext. 24160
Services FAX (905) 528-3773
McMaster University A. N. Bourns Bldg. Rm 131C
Hamilton, Canada L8S 4M1


From: Andrew Conn <>
Date: Tue, 24 Jan 1995 11:36:34 -0500
Subject: Release 2.0 of CUTE

Release 2.0 of CUTE is now available

CUTE (Constrained and Unconstrained Testing Environment) is a set of Fortran
subroutines, system tools and test problems in the area of nonlinear
optimization and nonlinear equations. Its purpose is to
- provide a way to explore an extensive collection of problems (over 700
different test problems today),
- provide means of comparing existing packages,
- provide a way to use a large test problem collection with new packages,
- provide a mechanism to manage and update the system efficiently, and
- do all the above on a variety of popular platforms.
The version 2.0 is that corresponding to the paper describing CUTE to appear
in Transactions of the AMS on Mathematical Software.

NEW * The new release provides interfaces to new optimization packages.
* In addition to the existing interfaces for MINOS, NPSOL, OSL, TENMIN,
* UNCMIN, VA15, VF13, VE09 and VE14, CUTE now contains interfaces for
* - COBYLA (MJD Powell derivative free constrained code)

NEW * The new release also supports two new computer platforms. Besides
* CUTE is now available with fully automated installation procedures for
* MSDOS (with the WATCOM Fortran compiler).

NEW * - The Fortran tools now provide "finite-elements" matrix format for the
* second derivatives.
* - Some driver routines have been improved.
* - Known bugs in the tools and test problems have also been corrected.

CUTE has been written by I. Bongartz, A.R. Conn (both at IBM, Watson Research
Center), Nick Gould (RAL, UK) and Ph. Toint (FUNDP, Belgium). A LaTeX
manuscript detailing the package may be obtained by email from any of the
authors. It is also included in the distribution.

CUTE 2.0 is written is standard ANSI Fortran 77. Single and double precision
versions are available. Machine dependencies are carefully isolated and
easily adaptable.

The package may be obtained in one of two ways. Firstly, the reader can
obtain CUTE electronically (and free of charge) via an anonymous ftp call to
the account at the Rutherford Appleton Laboratory joyous-gard
(Internet i.d., in the directory pub/cute), or at Facultes
Universitaires Notre-Dame de la Paix (Namur) (Internet
i.d., in the directory cute). We request that the userid is given
as the password. This will serve to identify those who have obtained a copy
via ftp.
In addition, there is a CUTE www page on

Secondly, the package can be obtained on a floppy disk or magnetic tape at a
minimal price, intended to recoup the costs of media, packaging, preparation
and courier delivery. Potentially interested parties should contact Ph. Toint
to obtain a suitable order form.

Ingrid Bongartz
Andy Conn
Nick Gould
Philippe Toint

January 19th, 1995.


From: Ken Jackson <>
Date: Tue, 24 Jan 1995 13:50:32 -0500
Subject: Report Distribution from University of Toronto

We plan to change the way we distribute reports. In the past, we
mailed reports to all those on our technical reports mailing list as
reports were issued. As our mailing list has grown and postage and
photocopying charges have increased, this has become quite expensive.
Moreover, we believe that many people who receive these reports may not
actually want them.

Advances in technology now allow us to adopt a better report
distribution scheme. If we produce an official departmental technical
report, we will mail it to the libraries on our reports list, but we
normally will not mail unsolicitated reports to individuals. Moreover,
many papers will not be made into official departmental technical
reports, but will be available in electronic form in our reports

Our new plan is to periodically e-mail a list of the titles of recent
reports to interested people with instructions on how to obtain any
report they wish to obtain by anonymous ftp. If you recently received
such a note from me, then your name is already on our reports mailing
list, so you need not reply to this message unless you want your name
removed from our mailing list. If you didn't receive such a note from
me, but you wish to have your name included on our mailing list, please
send me ( an e-mail message requesting that I add
your name to our list. Please include your complete snail-mail address
as well in case we ever need it.

Ken Jackson
for the
Scientific Computing Group,
Computer Science Department,
University of Toronto.


From: Yin Zhang <>
Date: Tue, 24 Jan 1995 17:28:25 -0400
Subject: LIPSOL beta-2.1 Release

LIPSOL Beta-2.1
Release Announcement
January 24, 1995

LIPSOL (Linear programming Interior-Point SOLvers) is a package that
uses Matlab's sparse-matrix data structure and MEX utilities to
achieve both programming simplicity and computational efficiency.

--- What's New? ---

(1) All sources for MEX-files included
Release Beta-2.1 includes all the sources for the LIPSOL
MEX-files: namely, the newly released version 0.2 of the ORNL
sparse Cholesky package by Dr. Esmond Ng and Dr. Barry Peyton
(along with Dr. Joseph Liu's multiple minimum degree ordering
routines), and our MEX interface routines.

(2) Performance improved
Though not a major update, beta-2.1 is generally faster than
beta-2. For example, under the default setting it solves PILOT87
in 38 iterations and 485 cpu-seconds on an R4400/150MHz CPU
(-mips2 flag used in fmex). But don't try fit2p unless you have
a LOT of memory and patience.

(3) "Memory leak problem" fixed
The false alarms were caused by the Matlab function SPARSE.

--- What platforms are supported? ---

Release beta-2.1 still contains one binary file for the MPS reader.
It supports: DEC (Ultrix 4.x), SGI (IRIX 5.2), Sun Sparcs (SunOS
4.1.3) and HP (HP-UX 9.0x) workstations. Of course there is no
guarantee that the binary would run on every machine from the above

-- Where to get LIPSOL? --

LIPSOL Beta-2.1 distribution consists of only one file for all
platforms. It can be retrieved through anonymous FTP from:
(also available from the WWW URL:
See the README files for information on installing and running
LIPSOL beta2.1.

Comments, suggestions and bug reports are most welcome!

Yin Zhang
Dept. of Mathematics and Statistics Email:
Univ. of Maryland Baltimore County Phone: (410) 455--3298
Baltimore, Maryland 21228-5398, USA Fax : (410) 455--1066


From: Heinz W. Engl <>
Date: Wed, 25 Jan 1995 12:28:16 EST
Subject: Survey and Bibliography on ABS Methods

ABS Methods

I refer to the announcement of the second international conference in
ABS methods in the last digest.
If you want to know more about ABS methods, read the survey article
by E.Spedicato that appeared in the Springer (Vienna) journal
Survey on Mathematics for Industry (Vol.3, No.3, 1993, pp. 217-232).
That article also contains more than 200 references.

Heinz W. Engl, Linz, Austria

Prof.Dr.Heinz W. Engl E-Mail:
Industriemathematik or
Institut fuer Mathematik secretary:
Johannes-Kepler-Universitaet Phone: +43-(0)732-2468; ext.9219 or 693,
Altenbergerstrasse 69 secretary: 9220; home: +43-(0)732-245518
A-4040 Linz Fax: +43-(0)732-2468855
Oesterreich / Austria Telex: 2-2323 uni li a


From: Anthony Fiacco <>
Date: Mon, 23 Jan 1995 23:51:47 -0500 (EST)
Subject: Comments on Fiacco-McCormick Book

Dear Colleague:

SIAM is planning a second printing of the 1990 edition of the 1968
Fiacco-McCormick book "Sequential Unconstrained Minimization Techniques",
which appeared in their classics in Applied Mathematics series. We want
to make the new edition as error-free as possible. If you have noted any
errors, inconsistencies, or ambiguities, either technical or grammatical,
would you kindly convey them to me. Any constructive remarks would also
valuable. Please send your comments to me by mail, fax (202-994-0245) or
by e-mail ( I would greatly appreciate receiving
these by 14 February 1995, and I will be most grateful for your help.

Anthony V. Fiacco


From: Julia Olkin <>
Date: Wed, 25 Jan 95 14:15:27 PST
Subject: Discussion and Linear Algebra Net

The Editors of the Linear Algebra digest, or LAnet, are starting a new
monthly feature. We will have a Monthly Question, whose purpose is to
stimulate discussions among the linear algebra aficionados. As such,
we have purposely chosen CONTROVERSIAL topics to generate lively (but
friendly) discussions. We encourage you to write us your opinion on
the question, and to submit questions of your own for consideration.
We do NOT intend to denigrate or insult anyone, and the questions we
pose do not necessarily reflect our own thinking. We are including
our first question here in the NAdigest. However, to receive future
questions and responses you will need to join the LAnet, by sending
your name and email address to

If you would like to answer this First Question,
please respond DIRECTLY to either of the two editors at: (Julia Olkin) (Juan Meza)
instead of sending your response to
We will collect all the responses and include them together in one
issue before bringing out the next Monthly Question. You may remain
anonymous but we would much prefer that you agree to include your

FIRST MONTHLY QUESTION: Is Numerical Linear Algebra dead? If not,
then what are the open questions?


From: Danny Hershkowitz <>
Date: Tue, 24 Jan 1995 08:45:13 +0200 (EET)
Subject: Changes in ILAS Information Center


As of December 28, 1994, we have shifted the IIC database to a new
machine. Consequently, we are happy to announce the following
significant improvement in the services offered by IIC.

There are now THREE ways to use the services of IIC:


Issue the command


where command stands for mosaic, xmosaic, netscape or lynx .


1. Enter 'ftp'
2. Enter 'anonymous' for the "user:" prompt
3. type your local userid as a PASSWORD
4. Once you get the first panel, type 'cd iic' to be connected to IIC
5. Issue any FTP command (type 'help' for available FTP commands)

III. E-mail

Include the command


(where "filename" is the file you request)

in the mail body of a message sent to .
The file will be sent to you in a mail file.
The message can include more than one request.

The list of files that can be obtained from IIC is included in the
file index.html .

Note the system is case sensitive and that all file names in IIC
(with the exception of index.html) are given in upper case


From: Trini Flores <>
Date: Fri, 27 Jan 95 09:50:19 EST
Subject: SIAM Annual Meeting


1995 SIAM Annual Meeting

October 23-26, 1995
Adam's Mark Hotel
Charlotte, North Carolina


The 1995 SIAM Annual Meeting will focus on computational science and
engineering -- its applications, and the interdisciplinary interactions that
yield new perspectives on research, education, and the contribution of
mathematics to science and industry.

The meeting will highlight the role of mathematics and computation in the
areas of control of large systems, environmental modeling, multi-disciplinary
design optimization, and biology and chemistry. The meeting will also
emphasize the role of numerical algorithms and software in computational
science and engineering. The ever-increasing role of computation in these
fields has led many institutions to install interdisciplinary degree programs
that focus on computational science.


Control of Large Systems
Some Results and Some Open Questions in the Control of Distributed
J.-L. Lions, College de France

Computational Science and Engineering Education
Experience with Electronic Textbooks for Computational Science and
Parallel Programming Education
Dennis Gannon, Indiana University, Bloomington

(Title to be determined)
Gene H. Golub, Stanford University

Computational Chemistry/Biology
Mathematics Problems Suggested by Computational Statistical Mechanics
Hans C. Andersen, Stanford University

The Topology of DNA
De Witt Sumners, Florida State University

Multidisciplinary Design Optimization
Industrial Strength Optimization
Paul D. Frank, Boeing Computer Services

Numerical Algorithms
Computing Fluid Flows in Complex Geometry
Marsha Berger, New York University

Environmental Modeling
Adaptive Projection Methods for Low Mach Number Flows
John B. Bell, Lawrence Livermore National Laboratory

Minisymposia are sessions consisting of four presentations on a single topic
that is consistent with the meeting themes. Following is a partial list of
minisymposia and organizers:

Finite Element Methods in Mechanics
Susanne C. Brenner, University of South Carolina, Columbia

Topology of Macromolecules
De Witt Sumners, Florida State University

Molecular Dynamics
Tamar Schlick, New York University

Computations in Image Reconstruction and Restoration
Robert J. Plemmons, Wake Forest University

Numerical Solution of Optimal Control Problems
Ekkehard W. Sachs, Universitat Trier, Germany

Multidisciplinary Design Optimization
Natalia Alexandrov, NASA Langley Research Center

Are You Interested in Organizing a Short Course?

SIAM short courses provide introductions to a timely topic in mathematics,
scientific computing and their applications. Short courses are typically one-
day long, and are designed to complement one of the meeting themes. Proposals
to organize a short course for this meeting, are subject to approval by the
organizing committee and SIAM. To obtain a short course proposal form and
guidelines, send your request by e-mail to

Proposals are due by April 14, 1995.

How to Contribute
You are invited to contribute a presentation at the SIAM annual meeting in one
or more of the following formats:

Contributed Presentations in Lecture or Poster Format
A lecture consists of a 12-minute presentation with an additional three
minutes for discussion. A poster presentation consists of the use of visual
aids, such as 8-1/2" x 11" sheets, mounted on a 4'x6' poster board. A poster
session is two hours long. Each contributor must submit a brief abstract not
exceeding seventy-five words on a SIAM abstract form.

Deadline for submission of abstracts: May 15, 1995.

Contributed Minisymposia
A minisymposium consists of four 25-minute presentations with an additional
five minutes for discussion. Prospective organizers are asked to submit a
proposal consisting of a title, a description (not exceeding one-hundred
words), and a list of speakers and titles of their presentations. Each
minisymposium speaker must submit a seventy-five word abstract. Proposals to
organize a minisymposium are encouraged, but not limited to be in line with
the main conference themes.

Deadline for submission of minisymposium proposals: April 14, 1995.

Electronic Mail Response and Submission
To obtain an abstract form and a minisymposium proposal form, and to submit
your abstracts and minisymposium proposals by electronic mail, send your
requests and submissions to Plain TeX or LaTeX macros are
available. You may continue to submit abstracts in hard copy form if you wish.
Mail it to SIAM, Attention: Conference Department, 3600 University City
Science Center, Philadelphia, PA 19104-2688.

Organizing Committee

Danny C. Sorensen (Chair)
Rice University

John E. Dennis, Jr.
Rice University

Carl T. Kelley
North Carolina State University

Barbara L. Keyfitz
University of Houston

Ahmed H. Sameh
University of Minnesota, Minneapolis

L. Ridgway Scott
University of Houston

Daniel D. Warner
Clemson University


From: Jeanne Butler <>
Date: Thu, 26 Jan 1995 10:26:12 -0400
Subject: Parallel Programming on the IBM SP2

The Cornell Theory Center will be offering the following
workshop for academic researchers:


March 28 - 31, 1995
Registration materials must be received by February 23, 1995.

This workshop will enable academic researchers to exploit the parallel
capabilities of the IBM Scalable POWERparallel System SP2.

The Cornell Theory Center's SP2 consists of 512 RISC processors connected
by a high performance switch. The size and power of this system allow
parallel programmers to decrease the wallclock time needed to run their
programs, or to attack larger problems by running over larger numbers of
nodes. The programming model is distributed memory.

The first three days of this lecture-lab workshop will cover concepts of
parallel programming, PVM (Parallel Virtual Machine) from Oak Ridge
National Laboratory, the IBM MPL message passing library, and additional
tools for development and analysis of parallel programs.


- SP2 Message Passing Library -- MPL
- Parallel Virtual Machine (PVM)
- Message Passing Interface (MPI) Overview - Performance Optimization
- Scaling to Exploit Parallelism
- Parallel Tools for Development/Debugging

The final day of the workshop will consist of special topics seminars.
Each of the special topics sessions will be limited to 10 people.

A detailed announcement can also be found online at

Jeanne Butler
Conference Assistant
Cornell Theory Center
427 Engineering and Theory Center Building
Ithaca, NY 14853-3801
(607) 254-8813
FAX: (607) 254-8888


From: Michael Ferris <>
Date: Mon, 23 Jan 95 10:50:08 -0600
Subject: Conference on Complementarity Problems


International Conference on Complementarity Problems:
Engineering & Economic Applications, and Computational Methods

November 1-4, 1995

Homewood Campus of The Johns Hopkins University, Baltimore, Maryland,

organized by:

Michael C. Ferris (, University of Wisconsin at Madison
Jong-Shi Pang (, The Johns Hopkins University

The conference will bring together, for the first time, engineers,
economists, industrialists, and academicians from the U.S. and abroad
who are involved in pure, applied, and/or computational research of
complementarity problems, to present and discuss the latest results in
this subject, and to offer suggestions for collaborative research and
further development of the field.

The conference will last for 4 days, consisting almost entirely of
invited presentations. There will be a small number of selective
contributed talks and the conference is limited to 100 participants
(including the speakers). A refereed volume of proceedings of the
conference will be published.

Conference Topics:

There are three major themes of the conference: engineering
applications, economic equilibria, and computational methods. Each
theme will be represented by experts in the area. Topics to be
covered in the conference are listed below.

Engineering applications:

Contact mechanics problems, structural mechanics problem, nonlinear
obstacle problems, elastohydrodynamic lubrication problems, traffic
equilibrium problems.

Economic applications:

Applied general economic equilibrium, game--theoretic models, NEMS.

Computational methods:

Pivotal and path following methods, smoothing techniques,
quadratic programming based methods, interior point methods, and
projection/proximal based methods; software development, modeling
language interfaces.

Contact one of the organizers for further details if you are interested
in participating at the conference or in contributing a paper for
possible presentation.


Michael C. Ferris Jong-Shi Pang
(on leave at)
Department of Economics Department of Mathematical Sciences
University of Colorado The Johns Hopkins University
Campus Box 256
Boulder, Colorado 80309 Baltimore, Maryland 21218
(303) 492-2651 (410) 516-7216
Email: Email:


From: Danny Hershkowitz <>
Date: Tue, 24 Jan 1995 08:38:51 +0200 (EET)
Subject: Haifa Matrix Theory Conference

MAY 29 - JUNE 1, 1995
Second Announcement

The Mathematics Department at the Technion, supported by its Institute
for Advanced Studies in Mathematics, will hold the Ninth Haifa Matrix
Theory Conference on May 29 - June 1, 1995.

The program will consist of plenary talks (30 minutes each) as well as
contributed talks (20 minutes each), aiming to cover all aspects of
matrix theory and linear algebra.

The members of the Organizing Committee - Avi Berman, Moshe Goldberg,
Danny Hershkowitz, Leonid Lerer, Raphael Loewy, and Abraham Zaks - invite
those who are interested in attending should request a registration form
from <> as soon as possible, and no later
than March 31, 1995. Those who wish to contribute a talk should attach a
short abstract.

Organizing Committee
Ninth Haifa Matrix Theory Conference
Department of Mathematics
Technion - Israel Institute of Technology
Haifa 32000, ISRAEL


From: Zbigniew Sikora <>
Date: Wed, 25 Jan 1995 14:30:38 +0100 (MET)
Subject: Workshop in Poland in Homogenization Theories


International Workshop on HOMOGENIZATION, THEORY OF
MIGRATION AND GRANULAR BODIES will be held from May 14th to May 17th 1995
in Mierki near Olsztyn in Poland.
Some 50 scientists are expected to participate after the registration.
Some general topics for the discussion:
-The concept of a mixture, averaging and homogenization theories
(kinematics, balance equations,...),
-Homogenization and localization,
-Constitutive theories for granular media,
-Near stationary flow and pulsating flow,
-Critical point phenomena (universal phenomena, renormalization),
-Local or global point attractors vs. strange attractors,
-Selforganization or chaos.

The deadlines are already exceeded but for some interesting papers we
have a possibility to include in proceedings of the Workshop.
We would be happy if we have a final version of the paper 15 March '95
before. Please contact us if you intend to join us as quickly as possible.

For additional information please contact me under direct address:
Please add the word 'KORMORAN'.

Sincerely, ZBIG.
Zbigniew Sikora, Technical University of Gdansk, Poland.


From: Bart Motmans <>
Date: Wed, 25 Jan 1995 14:36:28 +0100
Subject: Benelux Meeting on Systems and Control

The Organizing Committee has pleasure in inviting you to the
14-th Benelux Meeting on Systems and Control, to be held in
Houthalen, Belgium on March 29-31,1995 .

Scientific Program Overview

1)Invited Plenary Lectures

-Subspace Algorithms in Systems Identification and Telecommunication
Applications (Parts I and II) -
Prof. B. Ottersten (Royal Institute of Technology, Stockholm, Sweden)

-Adaptive Control, have we learned anything yet ? (parts I and II) -
Dr. R.L. Kosut (Integrated Systems Inc., Santa Clara, U.S.A. and
Stanford University, Stanford, USA)

2)Mini courses

-Convex Optimization and Control -
Prof. S. Boyd (Stanford University, Stanford, USA)

-Numerical Linear Algebra for Control -
Prof. P. Van Dooren (UCL, Louvain-la-Neuve, Belgium)

3)Contributed short lectures (142 abstracts)

Conference location

The meeting takes place in the conference center `Hengelhoef' in Houthalen,
Limburg, situated in quiet surroundings (300 ha of woods and ponds)
with nice walking possibilities. Accomodation is provided in the conference
center. The center has a large number of lecture rooms, as well as a bar,
and recreation and sports facilities, including tennis-courts (bring
your own racket), volleyball-courts, a subtropical swimming pool (free
entrance during the meeting). The atmosphere facilitates informal contacts.


The complete program program and the registration form is available by
anonymous ftp from '' in the directory
'pub/SISTA/publications/ben95_***.ps.Z' .

Registration is possible until Feb. 20, 1995.

Bart Motmans
Dept. of Electrical Engineering (ESAT)-SISTA
Kardinaal Mercierlaan 94
B-3001 Leuven
Tel. +32 (0) 16-321804 (office), +32 (0) 16-321709 (Secretary)
E-mail :


From: Goetz Alefeld <>
Date: Wed, 25 Jan 95 17:58:55 +0100
Subject: Error Bounds for Numerical Algorithms


-- Algorithms, Computer Arithmetic, Theory and Applications --


July 9--12, 1995, Universitaet Oldenburg, Fachbereich Mathematik, Germany
Organizing Chairmen: G. Alefeld, Karlsruhe and J. Herzberger, Oldenburg
supported by
International Association for Mathematics and Computers in Simulation (IMACS)
Gesellschaft fuer Angewandte Mathematik und Mechanik (GAMM)
Institut fuer Angewandte Mathematik der Universitaet Karlsruhe, Germany
Fachbereich Mathematik der Universitaet Oldenburg, Germany


Symposium Themes: Error-Bounds for Numerical Algorithms -- Computer
Arithmetic -- Two-Sided Methods -- Interval Arithmetic Algorithms --
Self-Validating Numerical Methods -- Computer Algebraic Methods in Interval
Arithmetic -- Automatic Differentiation with Error-Bounds

Purpose: This symposium is thought to be a satellite conference to the
ICIAM-World Congress in Hamburg and provides a forum for researchers and
appliers in the field of numerical algorithms with emphasis on the above
mentioned symposium themes. There will be invited lectures as well as short
communications (20 min.).

Call for Contributions: Potential contributers should submit abstracts
(one page) to the local organizer of the symposium (Prof. J.
Herzberger) not later than March 1, 1995.

Proceedings: It is planned to publish selected papers of the symposium
after a refereeing process in a proper form.

Symposium Venue: The symposium will be held at the Universitaet
Oldenburg (Wechloy Campus) in the city of Oldenburg. Oldenburg is located
about 150 km south-west of Hamburg. It has a main railway station and
moderately priced hotel accommodations in the city. There is a public bus
(line 6) directly to the campus.

Duration: The symposium will start on Sunday, July 9, in the early
afternoon and end on Wednesday, July 12, at noon.

If you want to receive further information, please contact

Prof. Dr. J. Herzberger
Fachbereich Mathematik
Universitaet Oldenburg
D - 26111 Oldenburg


From: Jean-Michel Muller <>
Date: Thu, 26 Jan 95 17:43:55 +0100
Subject: Computing with Real Numbers

Theoretical Computer Science
Special issue : Computing with Real Numbers
Call For Papers

Guest Editors: Jean-Claude Bajard, Christiane Frougny, Jean-Michel
Muller and Gilles Villard

Important Dates:

- Deadline for submissions of manuscripts (full papers only: extended
abstracts will be rejected): May 15, 1995

- Notification of acceptance/rejection: October 15, 1995

- Deadline for reception of final papers: December 1st, 1995

Efficient manipulation of (some) real numbers in computers is a new
frontier! Many interesting theoretical and algorithmic problems are
linked with that topic, and belong to quite distant fields such as
computer science, number theory, numerical analysis, computer algebra
and logics. The aim of this special issue of TCS is to present the
state of art in this increasing new domain, gathering contributions
coming from these different aeras. Researchers working on questions
related to real computer arithmetic are encouraged to present the
theoretical or algorithmic aspects of their results.

send 4 copies of a full paper to:

Jean-Michel Muller
Laboratoire LIP,
Ecole Normale Superieure de Lyon,
46 Allee d'Italie,
69364 LYON Cedex 07, FRANCE
(Phone +33 72 72 82 29 Fax +33 72 72 80 80)

Or a PostScript or LaTeX version of your full paper to:
(subject: TCS)

Before May 15. Please clearly indicate that your submission is to the
special issue of TCS.

Electronic submissions are encouraged, but please make sure that your
postScript files can be printed.

Language: English

IMPORTANT: A Conference devoted to the manipulation of real numbers
will be held in Saint-Etienne, France, in April 1995. Information on
that conference can be obtained by sending an electronic mail or by
writing to Jean-Michel Muller, at the address given above.


From: Panos Pardalos <>
Date: Thu, 26 Jan 95 20:12:18 EST
Subject: Symposium in Beijing on Operations Research

International Symposium on Operations Research
with Applications in Engineering, Technology,
and Management (ISORA)
Beijing, August 19-22, 1995
Sponsored by
the Asian-Pacific OR Center
within APORS and CAS

The symposium is intended to provide a forum for researchers working
in Operations Research, dealing with theoretical, computational, and
application aspect of optimization. Optimization is understood in the
widest sense including linear, nonlinear, stochastic, combinatorial,
and multiobjective systems. Papers presenting original research in
these areas are sought. Typical, but not exclusive, topics of
interest include:

- Linear and nonlinear programming
- Combinatorial and global optimization
- Multiobjective optimization
- Stochastic programming
- Scheduling and network flow
- Queuing systems
- Quality technology and reliability
- Simulation
- Optimizations in VLSI
- Neural network
- Financial modeling and analysis
- Manpower planning
- Production/Inventory control
- Flexible manufacturing systems
- Decision analysis
- Decision support systems
- Micro-computer software of OR methods

We especially appreciate papers on real-world applications.

Symposium Chair:
Xiang-Sun Zhang (Chinese Academy of Sciences, Beijing, China).

Symposium Advisors:
Guang-Hui Hsu (Chinese Academy of Sciences, China),
Masao Iri (Chuo University, Japan),
Santosh Kumar (RMIT, Australia),
Woong-Bae Rha(Korea),
Min-Yi Yue (Chinese Academy of Sciences, China).

International Program Committee Chair:
Ding-Zhu Du (University of Minnesota and Chinese Academy of Sciences)

Kan Cheng (Chinese Academy of Sciences, China),
Panos M. Pardalos (University of Florida, USA),
Dezhuang Yang (Chinese Academy of Sciences, China).

International Program Committee:
P.C. Bell (University of Western Ontario, Canada),
K.L. Chew (National University of Singapore),
C.A. Floudas (Princeton University, USA),
Jun Gu (University of Calgary, Canada),
T. Hasegawa (Kyoto University, Japan),
Gerd Infanger (Stanford University, USA),
N.K. Jaiswal (Institute for Systems Studies and Analysis, India),
Kin Keung Lai (City Polytechnic of Hong Kong),
S.J. Park (KAIST, Korea),
Huan-Wen Tang (Dalian University of Technology, China),
Henry Wolkowicz (University of Waterloo, Canada),
Cang-Pu Wu (Beijing Institute of Technology, China)
Gang Yu (University of Texas, USA),
Wenci Yu (East China University of Science and Technology, China),
Jenny Zhao (Silicon Graphics Computer System, USA),
Wei Zhao (Xi'an EUST, China),

Local Organizing Committee Chair:
Kan Cheng (Chinese Academy of Sciences, Beijing, China).

Authors are requested to submit 5 copies (in English) of an extended
abstract of at most ten pages by April 1, 1995 to one of the following

Professor Kan Cheng Dr. Ding-Zhu Du
Institute of Applied Mathematics Computer Science Department
Chinese Academy of Sciences University of Minnesota
Beijing 100080, P.R. China Minneapolis, MN 55455, U.S.A.

For information about program, registration and local arrangement,please
contact Kan Cheng at FAX 86-1-254-1689 or e-mail
or D.-Z. Du at FAX 1-612-625-0572 or e-mail

This conference is sponsored by:
The Asian-Pacific Operations Research Center
Co-sponsored by:
The Institute of Applied Mathematics, Chinese Academy of Sciences
The Operations Research Society of China
The National Natural Science Foundation of China


From: Ken Atkinson, Weimin Han, Florian Potra <>
Date: Fri, 27 Jan 1995 20:35:10 -0600
Subject: Announcement of Midwest NA Day

We are organizing a Midwest Numerical Analysis Day. It will be on
Saturday, April 29, and it will be held at the University of Iowa
in Iowa City, Iowa. There will be talks during both the morning
and afternoon, and we invite people to send a title and abstract
of proposed talks to us. We will try to accommodate as many as
possible among those who wish to speak.

Information concerning the conference will be stored in our
anonymous ftp site. Use
Then change to the subdirectory
We will store a list of some local motels, directions for getting
to the university, tentative programs, and other information in
the above anonymous ftp subdirectory. We expect there to be a small
registration fee to compensate for the cost of holding the meeting
(e.g., providing refreshments), payable upon arrival. We are
attempting to keep it as small as possible, tentatively $5.

If you intend on attending the meeting, then please so notify us
at the address
If you would like to give a talk, send a title to us, and if possible,
also an abstract. The deadline for submitting your title is March 31,
1995. Let us know if you have a preference about the time of your talk
(e.g., early morning). We will try to accommodate your request in
scheduling the talks. We will maintain an e-mail list of prospective
participants and will use it to send out a tentative program in early


From: Jacob White <>
Date: Wed, 25 Jan 95 10:27:13 EST
Subject: Position Available at MIT

The Research Laboratory of Electronics and the Computational
Hydrodynamics Facility at MIT has a postdoctoral position available in
developing adaptive numerical software for solving 3-D potential
problems. In particular, the research will focus on finding robust
approaches for adaptive gridding which efficiently fit with fast
integral formulation solvers based on the recently developed
fast-multipole and transform sparsification techniques. A variety of
application problems from electromagnetics, electromechanics, and
hydrodynamics will be used to determine approach and software
effectiveness. Candidates must have a PhD in mathematics, computer
science or in computationally oriented physical sciences or
engineering. A strong numerical analysis background and good
programming skills are required, and familiarity with potential theory,
adaptive gridding, boundary-element techniques, or fast-multipole algorithms
will be very helpful. Interested applicants should send a resume by
electronic mail to Professor Jacob White (


From: Lori Freitag <>
Date: Fri, 27 Jan 1995 13:01:12 -0600
Subject: Postdoctoral Position at Argonne National Laboratory

Postdoctoral Appointee
Mathematics and Computer Science Division
Argonne National Laboratory

The Mathematics and Computer Science Division of Argonne National
Laboratory invites applications for a postdoctoral research position.

The successful candidate will become part of a team effort aimed at
developing new parallel algorithms and software for problems arising
in unstructured mesh computation. Of particular interest are the
problems of mesh generation, adaptive refinement, mesh partitioning,
and the solution of the associated sparse linear systems. A possible
area of additional research is the investigation of virtual environment
visualization techniques for three-dimensional computational models
based on the unstructured mesh work. Part of the effort will include
work on specific finite-element combustion models. Information on the
parallel unstructured mesh computation project can be found at

We seek candidates with expertise in one or more of the following
areas: unstructured mesh algorithm and software design; finite-element
computation; and tool development for scientific computing applications,
particularly in parallel computing environments. Project members have
access to a state-of-the art computing and visualization environment,
including an IBM SP and a CAVE virtual environment system. Nominal
requirements include a Ph.D. in applied mathematics, computer science,
or engineering. Applicants must have received their Ph.D. not more than
three years prior to the beginning of the appointment.

Argonne is located in the southwestern Chicago suburbs, offering the
advantages of affordable housing and good schools, as well as easy access
to the cultural attractions of the city.

Applications should be addressed to Walter McFall, Box mcs-postdocx,
Employment and Placement, Argonne National Laboratory, 9700 S. Cass Avenue,
Argonne, IL 60439, and must include a resume and the names and addresses
of three references. For further information, contact Lori Freitag
(708-252-7246; or Paul Plassmann (708-252-4213;

Argonne is an affirmative action/equal opportunity employer.


From: SIAM <>
Date: Wed, 25 Jan 95 10:33:48 EST
Subject: Contents, SIAM Optimization

VOL.5,NO.2, 1995

Why Broyden's Nonsymmetric Method Terminates on Linear Equations
Dianne P. O'Leary

A New Infinity-Norm Path Following Algorithm for Linear Programming
Kurt M. Anstreicher and Robert A. Bosch

A Potential Reduction Algorithm with User-Specified Phase I-Phase II
Balance for Solving a Linear Program from an Infeasible Warm Start
Robert M. Freund

An Implicit Filtering Algorithm for Optimization of Functions with
Many Local Minima
P. Gilmore and C. T. Kelley

Indefinite Trust Region Subproblems and Nonsymmetric Eigenvalue Perturbations
Ronald J. Stern and Henry Wolkowicz

A Reduced Hessian Method for Large-Scale Constrained Optimization
Lorenz Biegler, Jorge Nocedal, and Claudia Schmid

A Robust Trust-Region Algorithm with a Nonmonotonic Penalty Parameter
Scheme for Constrained Optimization
Mahmoud El-Alem

A Class of Trust Region Methods for Nonlinear Optimization Problems
A. Sartenaer

Ladders for Travelling Salesmen
Sylvia C. Boyd, William H. Cunningham, Maurice Queyranne,
and Yaoguang Wang

On the Convergence of Fenchel Cutting Planes in Mixed-Integer Programming
E. Andrew Boyd

Subdifferential Convergence to Stochastic Programs
John R. Birge and Liqun Qi

Proximal Decomposition on the Graph of a Maximal Monotone Operator
Philippe Mahey, Said Oualibouch, and Pham Dinh Tao


End of NA Digest